Replaced drive from command line -- no swap partition?

[Z300M]

One of my pools was functioning with no reported errors although smartctl indicated Current_Pending_Sector
and Offline_Uncorrectable counts of 24 on one of the drives, so I decided to replace it anyway.

I couldn't find a way to do that from the GUI because the pool showed as Healthy and did not detect a problem drive with an option to replace it, so I replaced the drive from the command line:

zpool pool_name old_drive [by its gptid label] new_drive [by its /dev/daX id].

The pool is resilvering (four days to go), but when I execute

ls -l /dev/daX

it does not show two partitions, as the other drives in that pool do. Is this a problem? If so, how do I fix it?

 

[sretalla]

That's normal... there's a process you can use which includes partitioning the disk first, then using the second (data) partition gptid rather than the whole disk at the CLI.

If you have enough other disks with a swap partition, it's nothing to be concerned about.

The fix would be to offline, wipe and then replace it with itself using the GUI.

 

[danb35]

I couldn't find a way to do that from the GUI because the pool showed as Healthy and did not detect a problem drive with an option to replace it

If you couldn't find the option to replace the disk, you were looking in the wrong place (you should have been at the pool status page; the kebob menu for each disk in the pool gives you the options of edit, offline, and replace)--nothing in the process depends on FreeNAS deciding there's "a problem drive." There are a few problems with the procedure you followed:

• As you've already seen, the swap partition isn't there--not too big a deal by itself, but not optimal
• In fact, there's no partition table at all on the disk
• And by referring to the disk with its daX label, there's a (slight) risk of ZFS not recognizing it if that label should change.

The best way to fix it is to do what @sretalla says: offline the disk, wipe it, and replace it using the GUI. Failing that (for whatever reason), do it at the CLI properly--this resource includes instructions to do that:

Creating a degraded pool

WARNING: HERE BE DRAGONS. IF YOU DON'T KNOW EXACTLY WHAT YOU'RE DOING, DO NOT FOLLOW THESE INSTRUCTIONS. IN FACT, IF YOU NEED THESE INSTRUCTIONS, YOU PROBABLY SHOULDN'T FOLLOW THEM. There are a few cases where it may make sense to create a...

www.ixsystems.com

 

[Patrick M. Hausen]

If you use the command line like I prefer too, you need to perform all the steps. You should have partitioned the new drive and then
zpool replace pool gptid/oldid gptid/newid

Never add raw drives.

 

[Z300M]

If you couldn't find the option to replace the disk, you were looking in the wrong place (you should have been at the pool status page; the kebob menu for each disk in the pool gives you the options of edit, offline, and replace)--nothing in the process depends on FreeNAS deciding there's "a problem drive." There are a few problems with the procedure you followed:

• As you've already seen, the swap partition isn't there--not too big a deal by itself, but not optimal
• In fact, there's no partition table at all on the disk
• And by referring to the disk with its daX label, there's a (slight) risk of ZFS not recognizing it if that label should change.

The best way to fix it is to do what @sretalla says: offline the disk, wipe it, and replace it using the GUI. Failing that (for whatever reason), do it at the CLI properly--this resource includes instructions to do that:

Creating a degraded pool

WARNING: HERE BE DRAGONS. IF YOU DON'T KNOW EXACTLY WHAT YOU'RE DOING, DO NOT FOLLOW THESE INSTRUCTIONS. IN FACT, IF YOU NEED THESE INSTRUCTIONS, YOU PROBABLY SHOULDN'T FOLLOW THEM. There are a few cases where it may make sense to create a...

www.ixsystems.com

The pool status page (for both pools) is blank except for the column headings. If I click REFRESH, I get nothing but a white rectangle in the middle of the screen:

'Status:

Time Remaining: 268950 Seconds

Errors: 0

Date:

Name	Read	Write

'
(Beside each heading are up/down arrows that don't reproduce here, and neither do the Checksum and Status headings.)

I'm using FreeNAS-11.3-U4. I see there is a pending update, but the details did not seem to include anything that applies to my system.

 

[danb35]

The pool status page (for both pools) is blank except for the column headings.

That points to more serious problems--are you sure your boot device/pool are OK? And in this case, I guess it's up to you the sequence in which you want to address these problems--if you want to take care of the disk first, I'd still offline it, partition it properly, and replace it into the pool.

 

[Z300M]

That points to more serious problems--are you sure your boot device/pool are OK? And in this case, I guess it's up to you the sequence in which you want to address these problems--if you want to take care of the disk first, I'd still offline it, partition it properly, and replace it into the pool.

But the pools do show as HEALTHY: it's only when I click STATUS that the display misbehaves. And zpool status from the command line shows the pools as OK.

 

[Z300M]

But the pools do show as HEALTHY: it's only when I click STATUS that the display misbehaves. And zpool status from the command line shows the pools as OK.

And I have now discovered that this messed-up STATUS (non-)display occurs only with the Chrome browser, which I was using on my Linux machine to access the FreeNAS box: everything is fine when I use Firefox or Brave.